Effects of seismic soil-structure interaction on the analysis, design and behaviour of structures during their entire life cycle

Abstract

Dynamic response records of co-seismic pile performance are limited due to complexities and a lack of well-documented soil–pile response case histories. These limitations lead to inadequate provision of a standardised basis for the calibration and validation of the methods developed for seismic soil–pile superstructure interaction (SSPSI) and multi-hazard events problems. To address this, a series of numerical simulations (using finite element analysis (FEA)) for shaking table tests of scaled model piles in soft clay has been developed. The study identifies all numerical simulation aspects and soil constitutive criteria successfully. The shaking table test programme developed by Philip Meymand has been adopted as a physical test case. The study uses dimensional analysis to identify scale modelling criteria and develop a scaled soil and pile-supported structure model correctly. A unique numerical methodology is designed to permit multi-directional shear deformations, minimise boundary effects and replicate the free-field site response. Soil–structure interaction (SSI) effects, including the gap/slap mechanism and the consequences of kinematic and inertial force, are clearly shown. Full-scale co-seismic physical tests are complicated and even impossible as no fixed reference point is available as a benchmark. Most investigations performed after earthquake events analyse the consequences of the earthquake rather than the system behaviour. Employing a scaled testing technique, using shaking table tests in the one-g environment is a viable alternative. In this research, a calibration method for establishing the relationship between full-scale numerical analysis and scaled laboratory tests in one-g environment is developed. A sophisticated approach of scaling and validating full-scale seismic SSI problem is proposed. This considers the scaling concept of implied prototypes and ‘modelling of models’ techniques which can ensure a satisfactory level of accuracy. Pile integrity is commonly assessed during dynamic loading through simplified and uncodified analysis approaches. Two widely used seismic design codes (EC8 and ASCE) are compared, and revisions proposed. The effects of SSI on the seismic response of structures are determined. The findings reveal that structural response may exceed the codes’ limitations, making the provisions unsafe. The significance of the connection between the input motion and the site’s ground properties is also supported. The number of modes is associated with the response of the SSI model. The findings have important implications for the understanding of pile and pile group effects. Moreover, the definition of soil class F in EC8 and ASCE is ambiguous. The decision of designating class F or not for a project mostly depends on the experience of the personnel concerned. To reduce risk and to achieve a clear definition, the minimum thickness of sensitive clay to be considered to meet code condition for soil class F and the minimum thickness of sand layer that cuts off the continuity of soft clay layer to be no longer classified as F class are defined accurately. Furthermore, relatively little research has taken place into major multi-hazard events such as post-earthquake fire (PEF) and this is poorly covered in design codes. Structures subjected to an earthquake may experience partial damage, with an increased risk of structural failure during a later fire. A multi-hazard approach is developed, and two types of failure mechanisms are detected—global and local failure. The seismic SSI effects have been implicitly considered in the analyses. Finally, the study provides a robust evidence base for FEA aspect, including employing the correct soil model, in addition to an accurate scaling and validation methodology for co-seismic systems. It can contribute to a better understanding of seismic SSI codes provision including the application of SSI and the definition of soil class F, and delivers an effective methodology for multi-hazard analysis.